Science and validation
What is established, what is preliminary, what is open.
A model of health should be judged by the data it survives. Each claim below is marked with how far it has been tested.
The literature curves reproduce in a national sample
Of 35 catalogued hazard directions that NHANES measures, 29 reproduce with the same direction and a confidence interval excluding 1 (51,662 adults, 7,703 deaths). Three run opposite, the known reverse-causation pattern for LDL and non-HDL cholesterol in all-adult samples.
The landscape transports across populations
Hazards for inflammation, kidney function and glycaemia agree almost exactly between US adults and older Costa Ricans. Blood pressure and lipids bend in old age, which the model must express as age-dependent terrain.
People drift across it at measurable speeds
Within the same people after age 50, gait speed falls 0.82 SD per decade and grip strength 0.55 SD (HRS 2006–2022). Peak flow and gait speed accumulate the most mortality risk per decade of ageing.
Some single measurements are worth far more than others
Out of sample, grip strength, urine albumin/creatinine, RDW and the albumin/globulin ratio add the most to age and sex. Lipids add almost nothing within age bands.
One shared landscape with personal parameters
Whether shared curves plus a few individual parameters predict each person's path. This needs repeated blood measurements on the same people, now being requested from HRS, CLSA, ARIC and Alberta.
Rising variability as an early warning
Whether within-person fluctuation across systems signals loss of resilience before decline. A preregistered pilot in HRS is under way (osf.io/kzfs4).
The same hazards in three countries' data
Mortality hazard ratio per standard deviation, adjusted for age and sex, with 95% confidence intervals.
| Biomarker | NHANES · US, 20–84 | CRELES · Costa Rica, 60+ | HRS · US, 50+ |
|---|---|---|---|
| hs-CRP (log) | 1.30 1.26–1.33 | 1.30 1.19–1.43 | – |
| Serum creatinine | 1.25 1.22–1.28 | 1.27 1.16–1.39 | – |
| HbA1c | 1.18 1.16–1.21 | 1.20 1.10–1.31 | – |
| Resting heart rate | 1.21 1.19–1.24 | – | 1.20 1.17–1.23 |
| Waist circumference | 1.09 1.06–1.12 | 1.14 1.00–1.31 | 1.17 1.13–1.22 |
| Grip strength | 0.47 0.42–0.53 | – | 0.58 0.55–0.61 |
| Systolic BP | 1.08 1.05–1.10 | 0.89 0.78–1.00 | 1.06 1.04–1.09 |
Data
NHANES 1999–2018 with Linked Mortality to 2019; RAND Health and Retirement Study 1992–2022 with Tracker; ELSA HCAP 2018 and 2023; CRELES waves 1–3; MIDUS 3 biomarker project. Person-level cohort data are analysed only on a local machine. Only aggregate tables, with small cells suppressed, are published. AI tools help write and document code and never see individual records.